US 6732050 B2 Abstract A method, corresponding apparatus and corresponding system for determining a dynamical quantity of a receiver of signals conveying information useful in estimating the dynamical quantity, the method including the steps of: providing a single-point solution, by solving for the dynamical quantity of the receiver using a single-point solution having as an input the information useful in estimating the dynamical quantity being determined; providing a plurality of filter solutions (such as predictive filter solutions), and each assuming a different motion model for the receiver; and combining the plurality of filter solutions to provide a first value of the dynamical quantity based on weights that take into account the likelihood of the suitability of each motion model, with the likelihood determined on the basis of agreement of the first value of the dynamical quantity compared with a second value of the dynamical quantity as indicated by a single-point solution.
Claims(13) 1. A method for determining a dynamical quantity of a receiver of signals conveying information useful in estimating the dynamical quantity, the method comprising the steps of:
a) providing a single-point solution, by solving for the dynamical quantity of the receiver using a single-point solution having as an input the information useful in estimating the dynamical quantity being determined;
b) providing a plurality of filter solutions each assuming a different motion model for the receiver; and
c) combining the plurality of filter solutions to provide a first value of the dynamical quantity based on weights that take into account the likelihood of the suitability of each motion model, with the likelihood determined on the basis of agreement of the first value of the dynamical quantity compared with a second value of the dynamical quantity as indicated by a single-point solution.
2. The method of
3. The method of
4. An apparatus for determining a dynamical quantity of a host receiver of signals conveying information useful in estimating the dynamical quantity, the apparatus comprising:
a) means for providing a single-point solution, by solving for the dynamical quantity of the receiver using a single-point solution having as an input the information useful in estimating the dynamical quantity being determined;
b) means for providing a plurality of filter solutions each assuming a different motion model for the receiver; and
c) means for combining the plurality of filter solutions to provide a first value of the dynamical quantity based on weights that take into account the likelihood of the suitability of each motion model, with the likelihood determined on the basis of agreement of the first value of the dynamical quantity compared with a second value of the dynamical quantity as indicated by a single-point solution.
5. The apparatus of
6. The apparatus of
7. A system for determining a dynamical quantity of a component receiver of signals conveying information useful in estimating the dynamical quantity, the system comprising:
a receiver, including:
a) means for providing a single-point solution, by solving for the dynamical quantity of the receiver using a single-point solution having as an input the information useful in estimating the dynamical quantity being determined;
b) means for providing a plurality of filter solutions each assuming a different motion model for the receiver; and
c) means for combining the plurality of filter solutions to provide a first value of the dynamical quantity based on weights that take into account the likelihood of the suitability of each motion model, with the likelihood determined on the basis of agreement of the first value of the dynamical quantity compared with a second value of the dynamical quantity as indicated by a single-point solution.
8. The system of
9. The system of
10. The system of
11. The system of
12. The system of
13. The system of
Description The invention relates to satellite positioning systems, such as the global positioning system, and more particularly to algorithms used to determine the position of a satellite in such a system, using in some applications information provided by other (non-satellite-based) positioning systems (such as a cellular network positioning system) and sensors. In satellite positioning systems (such as the global positioning system or GPS), the position of a receiver (user) and its time offset from the system time (i.e. the correction to the receiver time at which the receiver is determined to be at the determined position) can be determined by using (pseudorange) measurements obtained from information (ephemerides and C/A-code phases) provided from at least four satellites. Such a determination can use satellite measurements at a particular instant of time, in what is called a single-point solution, a solution that in no way takes into account past information obtained from the satellites; any error in the measurements obtained from the satellites, including error from noise or multi-path, is reflected in such a single-point solution. Filtering with a Kalman filter (or some modification of such a filter) can instead be used to enhance the quality of the receiver's estimated track (by providing smoother, less noisy solutions than are provided by a single-point solution), and also to provide useable solutions in periods when satellite measurements are not available (because of poor signal conditions). The performance of any such filter is dependent on how the receiver's motion is modeled in the filter. Usually, instead of using an ordinary Kalman filter, what is called an extended Kalman filter (EKF), which is a linearized form of Kalman filter, is used, because a standard Kalman filter assumes that the measurement update equations are linear, and for positioning problems the measurement update equations, which involve the pseudoranges, are nonlinear. For a (standard) Kalman filter to be used, there has to be a linear relationship between the measurement vector m and state vector s, such that m=H·s, where H is some matrix. In GPS positioning, if the state vector is for example of the form [x y z t], where (x,y,z) indicates position and t represents clock bias, there is no such linear equation between pseudorange measurements and state. Instead, the i
which is obviously not a linear relationship. In an EKF, to be able to still use a Kalman type filter in an application where such a nonlinear relationship exists, the nonlinear relationship is approximated by a linear relationship by forming a truncated Taylor series of the nonlinear equation and taking the first, linear term of the series. In practice, this means that the H matrix in the equation m=H·s is approximated by the so-called Jacobian (known in the art) of the pseudorange equations. Thus, in an EKF, a standard Kalman filter (for linear systems) is applied to nonlinear systems (with additive white noise) by continually updating a linearization around a previous state estimate, starting with an initial guess. In other words, linear Taylor series approximation (no nonlinear terms) of the system function at the previous state estimate is made, and a linear Taylor series approximation of the observation function at the corresponding predicted position. Such an approach yields a relatively simple and efficient algorithm for handling a nonlinear model, but convergence to a reasonable estimate depends to a great extent on the-accuracy of the initial guess at the desired position; the algorithm may not converge if the initial guess is poor or if disturbances to the motion are so large that linearization is inadequate to describe the system. The prior art also teaches using what is called the interacting multiple model (IMM) solution, in which various motion models are assumed for the motion of the receiver (modules assuming slow turning, fast turning, slow accelerating, fast accelerating, and so on), and the outputs of the different models are combined based on weights that take into account how the predictions of the model agree with later measurements made on the basis of later information received from the satellites. In such an approach, each model (branch of the IMM solution) is implemented as an EKF. The Kalman filter solution (usually an EKF solution), as noted, is in principal superior to a single-point solution in that it uses more information and provides a correspondingly more educated receiver position estimate. The IMM solution is in principal suitable for a wider range of applications than any single-model solution. But the prior art of GPS teaches using only an EKF for each model of an IMM solution. The EKF is known to be inferior to an ordinary Kalman filter in performance, and the theoretical optimality results related to a Kalman filter do not apply to an EKF. In addition, using an EKF to implement each model of an IMM solution makes problematic using measurements based on information other than that provided by the positioning satellites, such as information from complementary positioning systems (e.g. cellular systems) and sensors including for example micro-electromechanical sensors such as inertial sensors (gyroscopes or accelerometers) and barometric altimeters. To integrate a measurement from a complementary positioning system into an EKF solution requires fusing another measurement to the state vector via some new nonlinear relationship which must be linearized and is therefore computationally costly. The prior art also teaches a two-stage solution, depicted in FIG. 1, in which a single-point solution (versus a Kalman filter type or predictive filter solution which takes into account past measurement points) is used based on pseudorange measurements, and the single-point solution is then provided to a single Kalman filter. However, such an approach does not take advantage of the wider range of applicability of a solution based on the IMM. What is needed, in order to provide a solution that is suitable in a wide range of applications and that provides reasonable estimates of a receiver's position even in poor signaling conditions, is an IMM positioning solution that does not suffer from the defects of one that uses an EKF to implement each model. Ideally, the sought-after IMM positioning solution would require less calculations than an EKF-based IMM solution, and would allow for using information from complementary positioning systems and sensors in a more straightforward manner. Accordingly, the present invention provides a method, a corresponding apparatus and a corresponding system for determining a dynamical quantity of a receiver of signals conveying information useful in estimating the dynamical quantity, the method including the steps of: providing a single-point solution, by solving for the dynamical quantity of the receiver using a single-point solution having as an input the information useful in estimating the dynamical quantity being determined; providing a plurality of filter solutions (such as predictive filter solutions) each assuming a different motion model for the receiver; and combining the plurality of filter solutions to provide a first value of the dynamical quantity based on weights that take into account the likelihood of the suitability of each motion model, with the likelihood determined on the basis of agreement of the first value of the dynamical quantity compared with a second value of the dynamical quantity as indicated by a single-point solution. In a further aspect of the invention, each predictive filter is an ordinary Kalman filter for which no linearizing of the measurement update equations of the Kalman filter is performed. In another further aspect of the invention, the dynamical quantity being determined is a quantity comprising one or more unknowns with respect to the receiver selected from the set consisting of: clock bias, position, clock drift, velocity, clock jerk, and acceleration. A system according to the invention, includes in some applications not only a receiver with components according to the invention, but also includes an external computing facility, not hosted by the receiver, and coupled to the receiver via wireless communications enabled by a cellular communication system, wherein at least some of the computation of either the single-point solution or the predictive filters is performed in the external computing facility and communicated to the receiver via wireless communication. In some other applications, a system, according to the invention, for determining a dynamical quantity of a receiver of signals conveying information useful in estimating the dynamical quantity includes not only a receiver with components according to the invention, but also at least one satellite for providing the signals conveying information useful in estimating the dynamical quantity. There are several advantages in the use of a two-stage IMM solution as opposed to a single-stage (EKF-based) IMM solution. First, the filters used are linear Kalman filters, and so are theoretically optimal and also have a predictable performance. One of the most significant problems associated with the EKF (whether or not used to implement the models of an IMM) is that the measurement equation linearization has to be carried out with respect to some particular solution point. In navigation applications, the particular point is the predicted position and time offset estimate. If the position estimate is badly mis-predicted, so is the linearization, leading to a diverging sequence of outputs from the EKF. A linear Kalman filter does not suffer from such a problem; as stated in A second advantage is that the two-stage approach of the invention allows using an extremely simple relationship between the measurement vector and the state vector, namely the identity relationship, s=m, since the “measurements” in the invention are actually state vector (single-point) solutions, so that the H matrix is the identity matrix, which simplifies all of the equations of the Kalman filters used in the invention, and many costly matrix operations are avoided, especially if sequential measurement processing is used. Altogether, the linearity and simplicity of the Kalman filters lead to substantial savings in computational costs. Another benefit worth mentioning is that with the invention it is possible to choose a first-stage pre-filter from among different single-point solution algorithms. For example, when only three satellites are seen, but the receiver knows its altitude, Phatak's point-solution, described in Yet another advantage is that in the present invention, an increase in the number of measurements does not significantly add to the cost in filtering computations. It is in general advantageous to use as many measurements as possible, but in case of a pure filter solution, the number of measurements increases the processing load of each filter. In the two-stage solution of the invention, however, the number of measurements affects only the first stage (the single-point solution), so that the cost of the second stage (the filters) is constant. Thus, instead of adding the computational load in all the parallel filters, the added information burdens the part of the algorithm that is performed only once at each filtering cycle. Finally, additional measurements from other positioning systems such as network-based systems or sensors such as inertial sensors can be readily used by the invention because of the simple form of the filter input “measurement” given by the first-stage single-point solution. The above and other objects, features and advantages of the invention will become apparent from a consideration of the subsequent detailed description presented in connection with accompanying drawings, in which: FIG. 1 is a block diagram of a two-stage, single-model filter apparatus according to the prior art; and FIG. 2 is a block diagram/flow diagram of a two-stage, IMM filter apparatus according to the invention. Referring now to FIG. 2, instead of using an extended Kalman filter (EKF) for filtering in an IMM solution, the invention uses the following two-stage IMM solution for the position, clock bias and possibly velocity and clock drift of a receiver (user). In the first stage, a single-point solution of the receiver's position and time offset is obtained (based on pseudorange measurements) using either a nonlinear closed form solution method or a linearized iterative point solution method. In the second stage, the output of the single-point solution is used as a measurement input to an IMM solution, one that can use any sort of predictive filter to implement each model, but preferably uses an ordinary Kalman filter. The second stage runs several preferably Kalman filters in parallel, each using a different motion model. A weighted combination of the estimates given by the different Kalman filters is formed at each filtering cycle. The weights are determined based on how well the estimates fit the measurements, i.e. based on the suitability of each of the models to express the motion of the receiver. The invention is not to be understood to be restricted to implementations in which all of the calculations of the receiver position are carried out by equipment housed in the receiver-itself, but also includes implementations in which some of the computation may be performed by external facilities in communication with the receiver via a cellular network. For the discussion here, however, it will be assumed that all of the computations are in fact carried out by equipment housed in the receiver itself. Still referring to FIG. 2, instead of raw pseudorange measurements being used as inputs to each model of the IMM solution, as in the prior art, the input to the IMM solution is the single point solution, i.e. the output of the first stage. The first stage is a nonlinear process (or a linearized process, if the iterative least squares method is used) that in effect pre-linearizes the measurements, resulting in a near sufficient statistic that summarizes the information contained in the raw pseudorange measurements (the measurements used as input to the single-point solution). A sufficient statistic is a mathematical data object that summarizes all the useful information contained in an entire set of data, to the extent that it can be used instead of the entire data set to provide the same result as would be obtained using the entire data set. In the present application, to say that the first stage results in a near sufficient statistic is to say that the computed point solution contains essentially the same information as a set of raw pseudorange measurements. This is important because if the point solution were not a near sufficient statistic and we used it as an input to the Kalman filters, then we might lose some information contained in the original pseudorange measurements. With such a first stage, the (second-stage) filters in the IMM can be linear, proper Kalman filters, i.e. standard Kalman filters; it is not necessary to linearize the measurement equations. As mentioned, the filters of a prior-art (single-stage) IMM would have to be EKFs, with all their disadvantages. It is also important to note that the first stage (single-point solution implementation) of a positioning apparatus according to the invention is common to all the parallel filters in the IMM. Thus, the first stage is performed only once per filtering cycle. The invention will now be described in more detail. The First Stage A suitable coordinate frame for an IMM in solving for a receiver's position in the vicinity of the earth's surface is the local East-North-Up (ENU) coordinate system. The East-North plane is defined to be tangent to the reference ellipsoid (approximating the surface of the earth) set out in the World Geodetic System 1984 (WGS-84), as explained for example in the textbook
where [e Equation set (1) can be solved either by an iterative linearization algorithm (least squares) or preferably by a nonlinear closed form algorithm. Several different closed form algorithms have been published. Most of them provide solutions only. in the determinate case (n=4). However, the earliest one, Bancroft's algorithm, provided in Principle of a Single Two-Stage Filter (Prior Art) In The filter of the second-stage assumes a certain motion model, according to which it estimates the receiver's dynamics. The model could be for example static (meaning that the filter assumes that the receiver is not moving at all), constant velocity or constant acceleration (or could be any other motion model). The receiver's state is formulated according to the model. For example, the state of a static filter is of the form x=[e n u Δt (Δt)] FIG. 1 also illustrates how measurements from other possible measurement systems (including position sensors) can be incorporated into a two-stage filter. Two-Stage IMM Filter (the Invention) Now referring to FIG. 2, the two-stage IMM consists of one first-stage single-point solution block (as in FIG. Discussion Thus, in the two-stage filtering of the invention, we first apply a single-point-solution method to the pseudoranges to get a single-point solution [x y z t] (or [e n u t] if the east-north-up coordinate system is used), and we use the single-point solution as the “measurement” m (as indicated in FIG. 2) given to each of the Kalman filters. Thus, the H-matrix of each of the Kalman filters is the identity matrix because the measurement vector is the state vector. So now we have a linear relationship between measurement and state, namely
If the state contains additional components such as clock drift dt (so that the state is [x y z t dt]), components that are not given by the single-point solution, we just add zero rows in the H matrix (which is otherwise then the identity matrix) so that we still have a linear relationship between measurement and state. Further, the same holds if we compute a single-point solution [dx dy dz dt] for the velocities dx, dy, dz and clock drift dt, using measured Doppler-phases. In that case, we combine the position single-point solution [x y z t] and the velocity single-point solution [dx dy dz dt] to get a “measurement” of the form m=[x y z t dx dy dz dt]. If we update the state vector in the form s=[x y z t dx dy dz dt] with such a measurement, we again have a relationship m=H·s in which H is the identity matrix. We can obtain acceleration from the velocity solutions in a similar extension. If we want to estimate a state including acceleration, and acceleration is not given by the single-point solution “measurements,” we just add zero rows to the H matrix, in places corresponding to the acceleration state components in the extended state vector s. (For example, if s=[x y z t dx dy dz dt d In other words, if we are interested in state vectors containing some of the components x, y, z, t, dx, dy, dz, dt (given in any convenient coordinate frame), by converting the pseudoranges (and possibly also Doppler-measurements) into single-point solutions before filtering, we obtain a linear relationship between the measurements and state, and get to use the linear Kalman filter without having to linearize the (Hmatrix) equation relating the state to the measurement. Embodiments in Which Velocity is Found Besides using the present invention in providing a position and clock bias solution via a Kalman filter-based IMM using pseudoranges as inputs to a first-stage single-point solution, the present invention is just as applicable to providing a velocity solution and clock drift solution. For a velocity and clock drift solution, instead of pseudoranges as inputs, the invention uses Doppler shift measurements (delta ranges) as inputs to a first-stage single-point solution. Moreover, the invention comprehends embodiments in which both velocity and position as well as clock drift are provided simultaneously, using both pseudoranges and delta ranges as inputs to a first-stage solution. In addition, by preprocessing delta-range measurements or pseudorange measurements at different times, it is possible to solve directly for acceleration (and also even clock jerk, the derivative of clock drift) using the preprocessed delta-range measurements or preprocessed pseudorange measurements as inputs to a first-stage “single-point” solution, the single point in time here being some point for which the preprocessing gives an approximate value of a measurement that would correspond to acceleration, i.e. a kind of delta-squared range measurement. Scope of the Invention It is to be understood that the above-described arrangements are only illustrative of the application of the principles of the present invention. In particular, it should be understood that as suggested above, the invention comprehends solving not only for the position of a receiver of signals from satellites of one or another global positioning systems, but also solving for any aspect of the dynamical state of such a receiver, such as the velocity or the acceleration, and doing so using any state-bearing information signal from any source, such as a signal conveying position information provided by a cellular network. In addition, the invention comprehends using other than ordinary Kalman filters or other predictive filters for the second stage; it comprehends using any kind of filter as a principal component of the second stage (the IMM stage) of the invention, although a predictive filter is preferred. Correspondingly, the output of the second stage is preferably a predicted value, but can also be a value corresponding to whatever other type of filter is used for the second stage. In the second stage, where the outputs of the different models are combined based on the suitability of each of the models, that suitability is preferably determined on the basis of measured pseudoranges or delta ranges (as indicated by the first stage solution), but could also be determined on the basis of other values. Also, the first stage of a solution according to the invention, i.e. the single-point solution, can be implemented as either a software algorithm or as hardware (sometimes called firmware). In addition, if the single-point solution is not available, the filter (second stage) could use some information from previous moments of time for calculation. Numerous other modifications and alternative arrangements may be devised by those skilled in the art without departing from the spirit and scope of the present invention, and the appended claims are intended to cover such modifications and arrangements. Patent Citations
Non-Patent Citations
Referenced by
Classifications
Legal Events
Rotate |